Top surface defects due to incorrect internal bridging

[scc]

**Edit for any new users:

• OrcaSlicer 2.0.0-beta fixes some, but not all of the incorrect g-code generation with the “Don’t filter out small internal bridges (beta)” option. Try limited filtering or no filtering (can generate a lot of wasted infill). This must be done before any later mitigations.
• For some shapes, especially very shallow (closer to 10 degrees) shapes, especially organic curves, slicer mitigations do not exist yet for all the difficulties in slicing 10-20 shallow slopes.
• For theses shapes, things can usually be solved by increasing thickness. Try using top thickness (e.g. 1.0mm → 1.8mm) instead of layers, because that will play better with variable layer height.
• If you are especially patient, you can reduce internal solid infill speeds to the 75-100% overhang speed.
• Lower temperature helps with some filaments, at the expense of worse layer adhesion.
  **

Hi,
I actually think I’ve got some of the root cause figured out for this specific print, but I thought I’d share some of what I learned and I still am looking for suggestions in case there is actually a “correct” fix for this. It seems like a slicer bug / “feature.”

I have printed 1000+ hours on my printers and am very familiar with the behavior of the PLA I use, but for the first time I started printing a very organic structures with variable layer height (at very high quality - heights range from 0.28 to 0.08mm). I started running into various versions of this pimply/orange skin defect, particularly in areas where the slope was very shallow:

I am aware that I set the seam in the wrong place for this test print - I’m not worried about that.

I tried a variety of things, so here’s what didn’t actually make a difference:

• Various advanced infill settings
• Infill type (no, gyroid doesn’t solve this at all, though it reduces nozzle banging)
• Top surface layers / thickness (mitigates it somewhat, but it’s an awful lot of extra material)
• Reducing MFR (I tried it anyway, but it’s actually very well tuned in for my material/nozzle/printer combination)
• Changing temperature (same as above, already tuned for these materials but tested anyway with no improvement)

Here’s the version with the fix:

What eventually made the difference was adjusting the internal solid infill speed down from the default 250mm/s to 170mm/s. For some reason, for very organic structures like this wedge, BambuSlicer fails to generate any bridge layers below the top layers. For example, in a standard cube, the very first layer that transitions between sparse infill to solid infill generates a bridge layer, which is printed at a much more appropriate 50mm/s. See slicer screenshot:

For the wedge I printed, it’s basically not printing a bridge transition layer, and instead bridging at the full 270mm/s which obviously results in very poor quality bridges and artifacts which consequently “telegraph” defects even through several top layers. At the shallowest layers where the internal solid infill layers look closest to being actual bridges and not overhangs, the defect is most pronounced.

Note that NO bridge lines are in fact used in the entire part.

(apologies for the slicer switch, for a variety of reasons I use both BambuSlicer and OrcaSlicer, but a standard cube does in fact slice the same in both bits of software)

Here’s what it looks like now - there are virtually no visible defects (I also moved the seam):

The small amounts of waviness in that area are almost certainly just telegraphed VFA, but well within what I’d tolerate for this. I might also just need to bump the solid infill speed lower, but of course that would impact print time pretty significantly (50mm/s will definitely work…but that’s a lot of solid infill time)

So my question is - is there some setting I’m missing that would help BambuSlicer or OrcaSlicer properly generate bridge layers for more organic top shapes like this one? Or is it more of a slicer bug/feature, in which case I might submit a feature/bug request on GitHub? The reduction in internal solid infill is really only a band-aid, since it slows down a lot of other solid infills that do not bridge and in fact do not need to be slowed down.

[Ed_3D_Shop]

Thank you for this topic!
I had this issue today at similar model. Maybe do you have another solution as to reduce the internal solid infill speed?
This was also my idea - reduce the speed for internal solid infill for critical layers… not really the best solution…
Best wishes,
Ed

[James_Marten]

Seems to be a known issue: Bumps on 10-20 degree gentle slopes · Issue #1653 · bambulab/BambuStudio · GitHub

Hope this will be fixed in Orca Slicer.

[ShaqFoo]

Have you tried using a modifier for the area that you need to slow down? I didn’t see you mention that anywhere in your post. A modifier would allow you to tweak the infill speed for only the area you need so as to not slow down the rest of the print. You could use the included modifer shapes to define the area but since your model is organic, just create your own and load it so more precisely modifies the area you want. I use both Orca and Bambu too. I use modifiers for speed changes and many other properties that are different from the global and object properties in the same print. You can get pretty granular with them. Good luck.

[frank.d]

Had the same problem (Orca, X1C, 0,2mm layers) with a planar surface at a shallow angle. Modifiers are a fix, but usually you find out after the print and it is time consuming.
Would be great, if they could add an option the the slicer settings reducing the speed of solid infill (%) under surfaces up the freely defined angle.
Unfortunately the issue has been closed on GitHub (Bambulab), with the explanations it only happens with small layer heights eg 0.08mm.

[dansan]

I seem to have had a similar issue, and reducing the speed did nothing. The issue was with the flow ratio. Try reducing this to 0.95 and see if this makes a difference. The calibration patches for this can be misleading, even if you are familiar with exactly how over-extrusion looks.

[Kodiacik]

Hi, I hlave problém with bottom layer of bridge.
The layer looks bad. I tried various filament.

[scc]

The picture I sent doesn’t actually show the full story. I still occasionally deal with this problem either by using OrcaSlicer and/or turning down the Internal Solid Infill speed to 50mm/s or lower.

There are actually two problems I’ve dealt with that are similar (particularly in large, low infill models). One happens when transitioning from infill to a top layer for a part with organic geometry on top. Ideally, the slicer should generate “real” bridges (like they should go from one edge of the infill to the next right under where the real top layers start), but for some reason Bambu Slicer’s infill bridge generation is broken. Orcaslicer does this correctly. I suspect this is a new feature since PrusaSlicer was also choking on the same geometry. I’ll find screenshots somewhere later.

This part is different. It’s both a 3D printing physics problem and maybe a slicer problem. With shallow geometries like this, very dense amounts of plastic can be deposited very close to each other, which might require slowing down due to higher cooling needs. This might be related to the physics of the issue rather than any slicer bug. It might need new rules, e.g. slow down when there is too much plastic concentrated in one area. You can see how dense it gets here:

Keep in mind that this is a “2 wall loop” print with 4 top shell layers @ 1mm, but in this shallow region the density has suddenly shot up to 10 loops, meaning the print has way less time to cool down for a very concentrated area.

It might be sort of a slicer problem though. Technically, I think the problem is that the “inner” (rightmost) line is actually either an overhang or bridge, since it has to bridge across empty air and/or it overhangs at a slight angle from the previous layer. However, when generating internal solid infill, neither bridge nor overhang angle rules are actually followed.

I may try writing some scripts to modify the gcode and slow down just those overhanging lines and see if that solves the problem. If it does, it could be possible to submit a feature request to OrcaSlicer and/or BambuSlicer for a slicer improvement. I have a feeling the OrcaSlicer team might be a little more aware of the problem given they have silently fixed the other infill bridge problem I found.

For the time being, the inefficient way of fixing this is two fold:

• Increasing the number of top layers will progressively “smooth” out any bridging or overhang artifacts from the poor transition layers. Since many smooth geometries will use variable layer height, I recommend setting “top shell thickness” rather than number of layers. You can go to 1mm or even higher depending on how bad it is.
• Decrease internal solid layer infill to 50mm/s or slower. This will make your Bambu feel a lot closer to an Ender, but it works and reduces the chance of you coming back to an ugly print. Usually you can guess which models need this based on the organic-ness of the top layer.
• Start with OrcaSlicer. For these types of defects, OrcaSlicer’s logic appears to be better than most other slicers at the moment.

[scc]

Well, what do you know. OrcaSlicer has been on top of this. PR#3319 addresses a large part of this problem, where bridging is incorrectly generated. While this does not necessarily fix the infill overhang speed issues, I’ve noticed it does significantly improve quality without needing nearly as many top layers or slowing down internal infill speeds.

I’m currently running a test print with one of my torture test models and I have noticed improvements.

This is available in the OrcaSlicer 2.0.0 beta.

[ChevaTomme]

I am so happy i found this post, while searching for solutions to internal bridging problems, it lead me to discover the settings to deactivate the filtering of small internal bridges in Orca!